Federated learning for predicting clinical outcomes in patients with COVID-19.

  title={Federated learning for predicting clinical outcomes in patients with COVID-19.},
  author={Ittai Dayan and Holger R. Roth and Aoxiao Zhong and Ahmed El Harouni and Amilcare Gentili and Anas Zainul Abidin and Andrew Liu and Anthony Beardsworth Costa and Bradford J. Wood and Chien-Sung Tsai and Chih-Hung Wang and Chun-Nan Hsu and C K Lee and Peiying Ruan and Daguang Xu and Dufan Wu and Eddie Huang and Felipe Campos Kitamura and Griffin Lacey and Gustavo C{\'e}sar de Ant{\^o}nio Corradi and Gustavo Nino and Hao-Hsin Shin and Hirofumi Obinata and Hui Ren and Jason C. Crane and Jesse Tetreault and Jiahui Guan and John W. Garrett and Joshua D. Kaggie and Jung Gil Park and Keith Dreyer and Krishna Juluru and Kristopher Kersten and Marcio Aloisio Bezerra Cavalcanti Rockenbach and Marius George Linguraru and Masoom A. Haider and Meena Abdelmaseeh and Nicola Rieke and Pablo F. Damasceno and Pedro M{\'a}rio Cruz e Silva and Pochuan Wang and Sheng Xu and Shuichi Kawano and Sira Sriswasdi and Soo Young Park and Thomas M. Grist and Varun Buch and Watsamon Jantarabenjakul and Weichung Wang and Won Young Tak and Xiang Li and Xihong Lin and Young Joon Kwon and Abood Quraini and Andrew Feng and Andrew N. Priest and Baris I Turkbey and Benjamin Scott Glicksberg and Bernardo Canedo Bizzo and Byung Seok Kim and Carlos Tor-D{\'i}ez and Chia-Cheng Lee and Chia-Jung Hsu and Chin Lin and Chiu-Ling Lai and Christopher P. Hess and Colin B. Compas and Deepeksha Bhatia and Eric Karl Oermann and Evan Leibovitz and Hisashi Sasaki and Hitoshi Mori and Isaac Yang and Jae Ho Sohn and Krishna Nand Keshava Murthy and Lijuan Fu and Matheus Ribeiro Furtado de Mendonça and Michael Fralick and Min Kyu Kang and Mohammad Adil and Natalie Gangai and Peerapon Vateekul and Pierre Elnajjar and Sarah E Hickman and Sharmila Majumdar and Shelley L. McLeod and Sheridan L Reed and Stefan Gr{\"a}f and Stephanie A. Harmon and Tatsuya Kodama and Thanyawee Puthanakit and Tony Mazzulli and Vitor Lima de Lavor and Yothin Rakvongthai and Yu Rim Lee and Yuhong Wen and Fiona J. Gilbert and Mona G. Flores and Quanzheng Li},
  journal={Nature medicine},
Federated learning (FL) is a method used for training artificial intelligence models with data from multiple sources while maintaining data anonymity, thus removing many barriers to data sharing. Here we used data from 20 institutes across the globe to train a FL model, called EXAM (electronic medical record (EMR) chest X-ray AI model), that predicts the future oxygen requirements of symptomatic patients with COVID-19 using inputs of vital signs, laboratory data and chest X-rays. EXAM achieved… 

Evaluation of federated learning variations for COVID-19 diagnosis using chest radiographs from 42 US and European hospitals

Federated Averaging (FedAvg) can significantly improve the generalization of the model compared to other personalization FL algorithms; however, at the cost of poor internal validity.

Towards Federated COVID-19 Vaccine Side Effect Prediction

An adaptive approach to fuse heterogeneous EHR data and apply data augmentation techniques working with a margin loss to overcome the data imbalance issue in the client model training is presented to address both challenges simultaneously in FedCovid.

Machine learning generalizability across healthcare settings: insights from multi-site COVID-19 screening

Three methods to achieve clinically-effective performances of COVID-19 diagnosis across four NHS Hospital Trusts are introduced and demonstrate that site-specific customization improves predictive performance when compared to other ready-made approaches.

Towards More Efficient Data Valuation in Healthcare Federated Learning using Ensembling

It is empirically show that SaFE computes values that are close to exact SVs, and that it performs better than current SV approximations.

Federated Learning Enables Big Data for Rare Cancer Boundary Detection

This study presentsings from the largest FL study to-date, involving data from 71 healthcare institutions across 6 continents, to generate an automatic tumor boundary detector for the rare disease of glioblastoma, utilizing the largest dataset of such patients ever used in the literature.

FLamby: Datasets and Benchmarks for Cross-Silo Federated Learning in Realistic Healthcare Settings

This work proposes a novel cross-silo dataset suite focused on healthcare, FLamby (Federated Learning AMple Benchmark of Your cross- silo strategies), to bridge the gap between theory and practice of cross-Silo FL.

FedDAR: Federated Domain-Aware Representation Learning

A novel method is proposed, FedDAR, which learns a domain shared representation and domain-wise personalized prediction heads in a decoupled manner for simplified linear regression settings, and it is theoretically proved that FedDar enjoys a linear convergence rate.

Using Real-Time Machine Learning to Prevent In-Hospital Severe Hypoglycemia: A prospective study

Evaluating the performance of a recently validated machine-learned model to predict inpatient hypoglycemia following its implementation into clinical care on cardiovascular and vascular surgery ward suggests that machine learning methods can be leveraged to prevent inpatient hyperglycemia.

MedPerf: Open Benchmarking Platform for Medical Artificial Intelligence using Federated Evaluation

MedPerf will enable federated evaluation in which models are securely distributed to different facilities for evaluation, thereby empowering healthcare organizations to assess and verify the performance of AI models in an efficient and human-supervised process, while prioritizing privacy.



Federated Learning of Electronic Health Records to Improve Mortality Prediction in Hospitalized Patients With COVID-19: Machine Learning Approach

The federated learning of COVID-19 electronic health record data shows promise in developing robust predictive models without compromising patient privacy.

Development and Validation of a Deep Learning Model for Prediction of Severe Outcomes in Suspected COVID-19 Infection

Results show that CO-RISK score achieved area under the Curve (AUC) of predicting MV/death (i.e severe outcomes) in 24 hours of 0.95, and 0.92 in 72 hours on the testing dataset, as well as compared to human performance.

Federated learning in medicine: facilitating multi-institutional collaborations without sharing patient data

It is shown that federated learning among 10 institutions results in models reaching 99% of the model quality achieved with centralized data, and the effects of data distribution across collaborating institutions on model quality and learning patterns are investigated.

Predicting COVID-19 mortality with electronic medical records

Age was found as the most important predictor of mortality in COVID-19 patients, and diseases that affect the pulmonary system, including interstitial lung disease, chronic obstructive pulmonary disease, lung cancer, and a smoking history, were important for predicting mortality.

The future of digital health with federated learning

This paper considers key factors contributing to this issue, explores how federated learning (FL) may provide a solution for the future of digital health and highlights the challenges and considerations that need to be addressed.

Federated Learning for Breast Density Classification: A Real-World Implementation

This study investigates the use of federated learning (FL) to build medical imaging classification models in a real-world collaborative setting and shows that despite substantial differences among the datasets from all sites and without centralizing data, it can successfully train AI models in federation.